Steady-state Free Precession Cine Research Articles

An artificial intelligence-based landmark model for early detection of aortic stenosis in Cardiac MRI

Abstract Background The utilisation of ratiometric intensity measures from cardiac magnetic resonance imaging (CMR) offers a promising biomarker for the diagnosis of aortic stenosis (AS)[1]. The Ao:LV ratio, an approximation of AS severity, is determined by comparing the relative blood signal intensity in the aorta and left ventricle (LV) within three-chamber (3Ch) cine sequences[1]. In patients with AS, the blood signal intensity of aortic blood above the valve is diminished compared to that within the LV in steady-state free precession (SSFP) 3Ch CMR cine. Although the Ao:LV ratio serves as a reliable indicator of AS severity, manual computation of this ratio may encounter difficulties due to visual obstructions caused by small LV cavities or prominent papillary muscles. Purpose This study aims to leverage artificial intelligence (AI) to automate the Ao:LV ratio analysis from 3Ch SSFP cine images and introduce the Ao:LA (aorta to left atrium) ratio, thus enhancing the diagnostic accuracy for AS severity. Methods Our methodology extends the ratiometric analysis to include both Ao:LV and Ao:LA ratios in three main steps: first, developing an AI-based algorithm for automated detection and tracking of cardiac landmarks; second, using these landmarks to automatically compute regions of interest (ROIs) in the cardiac chambers; and third, training and validating classification models with 5-fold cross-validation to select the most effective for AS severity classification. The final model's performance is assessed on a holdout test set. Data for model training and evaluation were sourced from CMR scans across three hospitals and two scanner types. Results We analysed 220 patients of which 78 patients had no AS and 122 had AS (mild AS, n = 29; moderate AS; n = 35, severe AS; n = 78). Our model yielded high efficacy in classifying AS, particularly in differentiating any grade of AS from a normal classification, which is clinically relevant. The Ao:LA ratio demonstrated a stronger correlation with AS severity class than Ao:LV, suggesting a more reliable biomarker. The proposed model demonstrated robust performance in AS classification, achieving an Area Under the Curve (AUC) of 0.845 in the automated binary classification of AS at any grade. The model exhibited a precision of 0.889, recall of 0.865, and an F1 score of 0.877 on a holdout set. Conclusion By the integration of AI in automating the analysis of CMR images through a novel ratiometric approach that includes Ao:LV and Ao:LA ratios, our model significantly improves the accuracy of early AS detection. This advancement enhances imaging strategies for AS evaluation, promoting more efficient and effective use of cardiac MRI sequences as a diagnostic tool for aortic stenosis. It has the potential to be implemented in real-time scanning protocols as a clinical decision support system.AI-inferred cardiac landmarks.ROC curve delineating AS prediction.

A pipeline for developing digital cardiac twins integrating cardiovascular magnetic resonance and electrocardiographic imaging: results from the MyoFit46 study

Abstract Background Personalised cardiovascular medicine uses multimodal screening and diagnostic tools, spanning from organ to tissue to molecular levels, to generate a complex and detailed patient phenotype. Cardiac digital twins are powerful tools capable of merging and explore these multimodal data by developing virtual, mechanistic and predictive versions of the patient's heart. State-of-the-art pipelines are mainly based on artificial intelligence and optimization process to replicate clinical data and not used it as an input. We present a proof-of-principle pipeline for building digital heart twins based on cardiovascular magnetic resonance (CMR) imaging and integrated electrocardiographic imaging (ECGI) for diverse clinical scenarios by integrating them into the model (Fig. 1). Methods CMR at 3T was performed on participants all born in the same week in March 1946, as part of the longest-running continued surveillance birth cohort: the National Survey of Health and Development. Cine steady-state free precession end-diastolic frames provided anatomical information per participant while CMR multiparametric mapping and late gadolinium enhancement (LGE) provided tissue characteristics to the electrical myocardial model used in the finite element solver Alya. Lastly, the ECGI activation map was used as a source to define the intrinsic endocardial activation sequence by applying a case-specific correction based on wall thickness and myofibers orientation. Activation-repolarization intervals map was used to introduce the personalised repolarization heterogeneities. Results In total, 406 prospectively recruited participants (77.8 ± 0.1 years, 44% male) had multiparametric CMR with ECGI for digital twin reconstruction. The pipeline’s performance was evaluated by measuring the root mean square error (RMSE) between the ECGI activation and repolarization maps with the in-silico ones obtained without applying the proposed pipeline (Baseline) and after its application (ecgi2model). Results obtained using ecgi2model pipeline presented a significantly reduced activation and repolarization RMSE (Fig. 2b), even providing accurate results in patients with different LGE patterns (Fig. 2c). Conclusion Personalised virtual heart models can be constructed at scale using real-world myocardial structural, functional, and electrical properties, from advanced CMR and ECGI data. Proposed digital twins provide a framework for personalised arrhythmic risk assessment based on real clinical data and also its applicability for in-silico clinical trials. Fig.1 Digital twins pipeline. Fig.2 a) Flattened epicardial map. b) RMSE of activation (ACT), repolarization (RPL) and activation-repolarization intervals (ARI) for the different methods methods. c) ACT and RPL maps obtained from ECGI, baseline simulation and ecgi2model simulation in a healthy participant with normal myocardial tissue, another with subendocardial LGE, and another with subepicardial LGE.Figure 1Figure 2

Deep learning super-resolution reconstruction for fast and high-quality cine cardiovascular magnetic resonance.

To compare standard-resolution balanced steady-state free precession (bSSFP) cine images with cine images acquired at low resolution but reconstructed with a deep learning (DL) super-resolution algorithm. Cine cardiovascular magnetic resonance (CMR) datasets (short-axis and 4-chamber views) were prospectively acquired in healthy volunteers and patients at normal (cineNR: 1.89 × 1.96 mm2, reconstructed at 1.04 × 1.04 mm2) and at a low-resolution (2.98 × 3.00 mm2, reconstructed at 1.04 × 1.04 mm2). Low-resolution images were reconstructed using compressed sensing DL denoising and resolution upscaling (cineDL). Left ventricular ejection fraction (LVEF), end-diastolic volume index (LVEDVi), and strain were assessed. Apparent signal-to-noise (aSNR) and contrast-to-noise ratios (aCNR) were calculated. Subjective image quality was assessed on a 5-point Likert scale. Student's paired t-test, Wilcoxon matched-pairs signed-rank-test, and intraclass correlation coefficient (ICC) were used for statistical analysis. Thirty participants were analyzed (37 ± 16 years; 20 healthy volunteers and 10 patients). Short-axis views whole-stack acquisition duration of cineDL was shorter than cineNR (57.5 ± 8.7 vs 98.7 ± 12.4 s; p < 0.0001). No differences were noted for: LVEF (59 ± 7 vs 59 ± 7%; ICC: 0.95 [95% confidence interval: 0.94, 0.99]; p = 0.17), LVEDVi (85.0 ± 13.5 vs 84.4 ± 13.7 mL/m2; ICC: 0.99 [0.98, 0.99]; p = 0.12), longitudinal strain (-19.5 ± 4.3 vs -19.8 ± 3.9%; ICC: 0.94 [0.88, 0.97]; p = 0.52), short-axis aSNR (81 ± 49 vs 69 ± 38; p = 0.32), aCNR (53 ± 31 vs 45 ± 27; p = 0.33), or subjective image quality (5.0 [IQR 4.9, 5.0] vs 5.0 [IQR 4.7, 5.0]; p = 0.99). Deep-learning reconstruction of cine images acquired at a lower spatial resolution led to a decrease in acquisition times of 42% with shorter breath-holds without affecting volumetric results or image quality. Question Cine CMR acquisitions are time-intensive and vulnerable to artifacts. Findings Low-resolution upscaled reconstructions using DL super-resolution decreased acquisition times by 35-42% without a significant difference in volumetric results or subjective image quality. Clinical relevance DL super-resolution reconstructions of bSSFP cine images acquired at a lower spatial resolution reduce acquisition times while preserving diagnostic accuracy, improving the clinical feasibility of cine imaging by decreasing breath hold duration.

Deep-Learning-Based Disease Classification in Patients Undergoing Cine Cardiac MRI.

Automated approaches may allow for fast, reproducible clinical assessment of cardiovascular diseases from MRI. To develop an MRI-based deep learning (DL) disease classification algorithm to distinguish among normal subjects (NORM), patients with dilated cardiomyopathy (DCM), hypertrophic cardiomyopathy (HCM), and ischemic heart disease (IHD). Retrospective. A total of 1337 subjects (55% female), comprising normal subjects (N = 568), and patients with DCM (N = 151), HCM (N = 177), and IHD (N = 441). Balanced steady-state free precession cine sequence at 1.5/3.0 T. Bi-ventricular morphological and functional features and global and segmental left ventricular strain features were automatically extracted from short- and long-axis cine images. Variational autoencoder models were trained on the extracted features and compared against consensus disease label provided by two expert readers (13 and 14 years of experience). Adding unlabeled, normal data to the training was explored to increase specificity of NORM class. Tenfold cross-validation for model development; mean, standard deviation (SD) for measurements; classification metrics: area under the curve (AUC), confusion matrix, accuracy, specificity, precision, recall; 95% confidence intervals; Mann-Whitney U test for significance. AUCs of 0.952 for NORM, 0.881 for DCM, 0.908 for HCM, and 0.856 for IHD and overall accuracy of 0.778 were obtained, with specificity of 0.908 for the NORM class using both SAX and LAX features. Longitudinal strain features slightly improved classification metrics by 0.001 to 0.03 points, except for HCM-AUC. Differences in accuracy, metrics for NORM class and HCM-AUC were statistically significant. Cotraining using unlabeled data increased the specificity for the NORM class to 0.961. Cardiac function features automatically extracted from cine MRI have potential to be used for disease classification, especially for normal-abnormal classification. Feature analyses showed that strain features were important for disease labeling. Cotraining using unlabeled data may help to increase specificity for normal-abnormal classification. 3 TECHNICAL EFFICACY: Stage 1.

The Association of Epicardial Adipose Tissue Volume and Atrial Fibrillation in Patients With Hypertrophic Cardiomyopathy: As Assessed by Cardiac MR.

Epicardial adipose tissue (EAT) is a metabolically active visceral fat linked to cardiovascular disease. Prior studies demonstrated the predictive value of EAT volume (EATV) in atrial fibrillation (AF) among hypertrophic obstructive cardiomyopathy patients. To investigate the association between EATV and AF in hypertrophic cardiomyopathy (HCM). Retrospective. Two hundred and twenty-four HCM patients (including 79 patients with AF and 145 patients without AF, 154 men) and 80 healthy controls (54 men). 3.0 T scanner; balanced steady-state free precession (SSFP) cine sequence, gradient echo. EAT thickness was assessed in the 4-chamber and basal short-axis planes. EAT volume was calculated by outlining the epicardial border and visceral pericardium layer on short-axis cine images. Shapiro-Wilk test, Student's t test or the Mann-Whitney U test, chi-square test or Fisher's exact test, Multivariate linear regression analyses, Multivariable binary logistic regression analysis. Intraclass correlation coefficient. Significance was determined at P < 0.05. EATV and EAT volume index (EATVI) were significantly greater in HCM patients with AF than those without AF (126.6 ± 25.9 mL vs. 90.5 ± 24.5 mL, and 73.0 ± 15.9 mL/m2 vs. 51.3 ± 13.4 mL/m2). EATVI was associated with AF in multivariable linear regression analysis among HCM patients (β = 0.62). Multivariable logistic regression analysis revealed that compared to other indicators, the area under curve (AUC) of EATVI was 0.86 (cut-off, 53.9 mL/m2, 95% CI, 0.80-0.89), provided a better performance, with the sensitivity of 96.2% and specificity of 58.6%. The combined model exhibited superior association with AF presence compared to the clinical model (AUC 0.96 vs. 0.76) and the imaging model (AUC 0.96 vs. 0.93). EATVI was associated with AF. EATVI was significantly correlated with incident AF, and provided a better performance in HCM patients compared to other indicators. 3 TECHNICAL EFFICACY: Stage 2.

Bayesian Uncertainty Estimation by Hamiltonian Monte Carlo: Applications to Cardiac MRI Segmentation

Deep learning (DL)-based methods have achieved state-of-the-art performance for many medical image segmentation tasks. Nevertheless, recent studies show that deep neural net- works (DNNs) can be miscalibrated and overconfident, leading to ”silent failures” that are risky for clinical applications. Bayesian DL provides an intuitive approach to DL failure de- tection, based on posterior probability estimation. However, the posterior is intractable for large medical image segmentation DNNs. To tackle this challenge, we propose a Bayesian learning framework using Hamiltonian Monte Carlo (HMC), tempered by cold posterior (CP) to accommodate medical data augmentation, named HMC-CP. For HMC compu- tation, we further propose a cyclical annealing strategy, capturing both local and global geometries of the posterior distribution, enabling highly efficient Bayesian DNN training with the same computational budget as training a single DNN. The resulting Bayesian DNN outputs an ensemble segmentation along with the segmentation uncertainty. We evaluate the proposed HMC-CP extensively on cardiac magnetic resonance image (MRI) segmentation, using in-domain steady-state free precession (SSFP) cine images as well as out-of-domain datasets of quantitative T1 and T2 mapping. Our results show that the proposed method improves both segmentation accuracy and uncertainty estimation for in- and out-of-domain data, compared with well-established baseline methods such as Monte Carlo Dropout and Deep Ensembles. Additionally, we establish a conceptual link between HMC and the commonly known stochastic gradient descent (SGD) and provide general insight into the uncertainty of DL. This uncertainty is implicitly encoded in the training dynamics but often overlooked. With reliable uncertainty estimation, our method provides a promising direction toward trustworthy DL in clinical applications. We release our code in &lt;a href='https://gitlab.tudelft.nl/yidongzhao/hmc_uncertainty'&gt;https://gitlab.tudelft.nl/yidongzhao/hmc_uncertainty&lt;/a&gt;

Incremental Prognostic Value of Left Atrial Strain in Patients With Suspected Myocarditis and Preserved Left Ventricular Ejection Fraction.

Analysis of left atrial (LA) strain and left atrioventricular coupling index (LACI) have prognostic value in cardiovascular diseases. However, the prognostic value of LA strain and LACI in patients with suspected myocarditis and preserved left ventricular ejection fraction (LVEF) is unclear. To investigate the prognostic value of LA strain and LACI in patients with suspected myocarditis and preserved LVEF in comparison with conventional MRI outcome predictors. Retrospective. One hundred sixty-five patients with clinically suspected myocarditis and preserved LVEF with available follow-up data. Steady-state free precession cine and phase-sensitive inversion recovery segmented gradient echo late gadolinium enhancement sequences at 3.0 T. Left ventricular (LV) and LA strain were evaluated using feature tracking. LACI was calculated as the ratio of LA and LV volumes at LV end-diastole. Patients were followed-up with the primary endpoint being major adverse cardiovascular events (MACE). Independent-samples t-test and Mann-Whitney U test to compare patients with and without MACE, receiver operating characteristic (ROC) curve analysis to define high/low risk groups, Kaplan-Meier survival analysis and Cox proportional hazards regression to assess prognosis. A P value of <0.05 was considered statistically significant. The associations of LV strain parameters (including global radial, circumferential, and longitudinal strain) and LACI with MACE were not significant (P = 0.511, 0.108, 0.148, and 0.847, respectively). An optimal LA conduit strain (Ԑe) cutoff value of 10.4% was identified to best classify patients into low- and high-risk groups. Only Ԑe was significantly associated with MACE in both univariable (hazards ratio [HR] 0.936, 95% confidence interval [CI] 0.884-0.991) and multivariable Cox survival analyses (HR 0.937, 95% CI 0.884-0.994). LA conduit strain has prognostic value in patients with suspected myocarditis and preserved LVEF, incremental to conventional MRI outcome predictors, whereas LACI was not associated with MACE occurrence. 3 TECHNICAL EFFICACY: Stage 2.

Evaluation of Left Ventricular Flow Kinetic Energy by Four-Dimensional Blood Flow MRI in Nondialysis Chronic Kidney Disease Patients.

Chronic kidney disease (CKD) is associated with increased, and early cardiovascular disease risk. Changes in hemodynamics within the left ventricle (LV) respond to cardiac remodeling. The LV hemodynamics in nondialysis CKD patients are not clearly understood. To use four-dimensional blood flow MRI (4D flow MRI) to explore changes in LV kinetic energy (KE) and the relationship between LV KE and LV remodeling in CKD patients. Retrospective. 98 predialysis CKD patients (Stage 3: n = 21, stage 4: n = 21, and stage 5: n = 56) and 16 age- and sex-matched healthy controls. 3.0 T/balanced steady-state free precession (SSFP) cine sequence, 4D flow MRI with a fast field echo sequence, T1 mapping with a modified Look-Locker SSFP sequence, and T2 mapping with a gradient recalled and spin echo sequence. Demographic characteristics (age, sex, height, weight, blood pressure, heart rate, aortic regurgitation, and mitral regurgitation) and laboratory data (eGFR, Creatinine, hemoglobin, ferritin, transferrin saturation, potassium, and carbon dioxide bonding capacity) were extracted from patient records. Myocardial T1, T2, LV ejection fraction, end diastolic volume (EDV), end systolic volume, LV flow components (direct flow, delayed ejection, retained inflow, and residual volume) and KE parameters (peak systolic, systolic, diastolic, peak E-wave, peak A-wave, E/A ratio, and global) were assessed. The KE parameters were normalized to EDV (KEiEDV). Parameters were compared between disease stage in CKD patients, and between CKD patients and healthy controls. Differences in clinical and imaging parameters between groups were compared using one-way ANOVA, Kruskal Walls and Mann-Whitney U tests, chi-square test, and Fisher's exact test. Pearson or Spearman's correlation coefficients and multiple linear regression analysis were used to compare the correlation between LV KE and other clinical and functional parameters. A P-value of <0.05 was considered significant. Compared with healthy controls, peak systolic (24.76 ± 5.40 μJ/mL vs. 31.86 ± 13.18 μJ/mL), systolic (11.62 ± 2.29 μJ/mL vs. 15.27 ± 5.10 μJ/mL), diastolic (7.95 ± 1.92 μJ/mL vs. 13.33 ± 5.15 μJ/mL), peak A-wave (15.95 ± 4.86 μJ/mL vs. 31.98 ± 14.51 μJ/mL), and global KEiEDV (9.40 ± 1.64 μJ/mL vs. 14.02 ± 4.14 μJ/mL) were significantly increased and the KEiEDV E/A ratio (1.16 ± 0.67 vs. 0.69 ± 0.53) was significantly decreased in CKD patients. As the CKD stage progressed, both diastolic KEiEDV (10.45 ± 4.30 μJ/mL vs. 12.28 ± 4.85 μJ/mL vs. 14.80 ± 5.06 μJ/mL) and peak E-wave KEiEDV (15.30 ± 7.06 μJ/mL vs. 14.69 ± 8.20 μJ/mL vs. 19.33 ± 8.29 μJ/mL) increased significantly. In multiple regression analysis, global KEiEDV (β* = 0.505; β* = 0.328), and proportion of direct flow (β* = -0.376; β* = -0.410) demonstrated an independent association with T1 and T2 times. 4D flow MRI-derived LV KE parameters show altered LV adaptations in CKD patients and correlate independently with T1 and T2 mapping that may represent myocardial fibrosis and edema. Stage 3.

Impact of Type 2 Diabetes Mellitus on Left Atrioventricular Coupling and Left Atrial Deformation in Patients with Essential Hypertension: An MRI Feature Tracking Study.

Hypertension (HTN) and type 2 diabetes mellitus (T2DM) are both associated with left ventricular (LV) and left atrial (LA) structural and functional abnormalities; however, the relationship between the left atrium and ventricle in this population is unclear. To identify differences between hypertensive patients with and without T2DM as the basis for further investigation the atrioventricular coupling relationship. Cross-sectional, retrospective study. 89 hypertensive patients without T2DM [HTN (T2DM-)] (age: 58.4 +/- 11.9 years, 48 male), 62 hypertensive patients with T2DM [HTN (T2DM+)] (age: 58.5 +/- 9.1 years, 32 male) and 70 matched controls (age: 55.0 +/- 9.6 years, 37 male). 2D balanced steady-state free precession cine sequence at 3.0 T. LA reservoir, conduit, and booster strain (εs, εe, and εa) and strain rate (SRs, SRe, and SRa), LV radial, circumferential and longitudinal peak strain (PS) and peak systolic strain rate and peak diastolic strain rate (PSSR and PDSR) were derived from LA and LV cine images and compared between groups. Chi-square or Fisher's exact test, one-way analysis of variance, analysis of covariance, Pearson's correlation, multivariable linear regression analysis, and intraclass correlation coefficient. A P value <0.05 was considered significant. Compared with controls, εs, εe, SRe and PS-longitudinal, PDSR-radial, and PDSR-longitudinal were significantly lower in HTN (T2DM-) group, and they were even lower in HTN (T2DM+) group than in both controls and HTN (T2DM-) group. SRs, εa, SRa, as well as PS-radial, PS-circumferential, PSSR-radial, and PSSR-circumferential were significantly lower in HTN (T2DM+) compared with controls. Multivariable regression analyses demonstrated that: T2DM and PS-circumferential and PS-longitudinal (β = -4.026, -0.486, and -0.670, respectively) were significantly associated with εs; T2DM and PDSR-radial and PDSR-circumferential were significantly associated with εe (β = -3.406, -3.352, and -6.290, respectively); T2DM and PDSR-radial were significantly associated with SRe (β = 0.371 and 0.270, respectively); T2DM and PDSR-longitudinal were significantly associated with εa (β = -1.831 and 5.215, respectively); and PDSR-longitudinal was significantly associated with SRa (β = 1.07). In hypertensive patients, there was severer LA dysfunction in those with coexisting T2DM, which may be associated with more severe LV dysfunction and suggests adverse atrioventricular coupling. 3. Stage 3.

Comparative analysis of left ventricle function and deformation imaging in short and long axis plane in cardiac magnetic resonance imaging.

Advancements in cardiac imaging have revolutionized our understanding of ventricular contraction. While ejection fraction (EF) is still the gold standard parameter to assess left ventricle (LV) function, strain imaging (SI) has provided valuable insights into ventricular mechanics. The lack of an integrative method including SI parameters in a single, validated formula may limit its use. Our aim was to compare different methods for evaluating global circumferential strain (GCS) and their relationship with global longitudinal strain (GLS) and EF in CMR and how the different evaluations fit in the theoretical relationship between EF and global strain. Retrospective monocenter study. Inclusion of every patient who underwent a CMR during a 15 months period with various clinical indication (congenital heart defect, myocarditis, cardiomyopathy). A minimum of three LV long-axis planes and a stack of short-axis slices covering the LV using classical steady-state free precession cine sequences. A single assessment of GLS on long axis (LAX) slices and a double assessment of GCS and EF with both short axis (SAX) and LAX slices were made by a single experienced CMR investigator. GCS-SAX and GCS-LAX were correlated (r = 0.77, P < 0.001) without being interchangeable with a high reproducibility for GCS, GLS and EF. EF calculated from LAX images showed an overestimation compared to EF derived from SAX images of 7%. The correlation between calculated EF and theoretical EF derived from SI was high (r = 0.88 with EF-SAX, 0.95 with EF-LAX). This study highlights the need to integrate strain imaging techniques into clinical by incorporating strain parameters into EF calculations, because it gives a deeper understanding of cardiac mechanics.

Fully Automated Valve Segmentation for Blood Flow Assessment From 4D Flow MRI Including Automated Cardiac Valve Tracking and Transvalvular Velocity Mapping.

Automated 4D flow MRI valvular flow quantification without time-consuming manual segmentation might improve workflow. Compare automated valve segmentation (AS) to manual (MS), and manually corrected automated segmentation (AMS), in corrected atrioventricular septum defect (c-AVSD) patients and healthy volunteers, for assessing net forward volume (NFV) and regurgitation fraction (RF). Retrospective. 27 c-AVSD patients (median, 23 years; interquartile range, 16-31 years) and 24 healthy volunteers (25 years; 12.5-36.5 years). Whole-heart 4D flow MRI and cine steady-state free precession at 3T. After automatic valve tracking, valve annuli were segmented on time-resolved reformatted trans-valvular velocity images by AS, MS, and AMS. NFV was calculated for all valves, and RF for right and left atrioventricular valves (RAVV and LAVV). NFV variation (standard deviation divided by mean NFV) and NFV differences (NFV difference of a valve vs. mean NFV of other valves) expressed internal NFV consistency. Comparisons between methods were assessed by Wilcoxon signed-rank tests, and intra/interobserver variability by intraclass correlation coefficients (ICCs). P < 0.05 was considered statistically significant, with multiple testing correction. AMS mean analysis time was significantly shorter compared with MS (5.3 ± 1.6 minutes vs. 9.1 ± 2.5 minutes). MS NFV variation (6.0%) was significantly smaller compared with AMS (6.3%), and AS (8.2%). Median NFV difference of RAVV, LAVV, PV, and AoV between segmentation methods ranged from -0.7-1.0 mL, -0.5-2.8 mL, -1.1-3.6 mL, and - 3.1--2.1 mL, respectively. Median RAVV and LAVV RF, between 7.1%-7.5% and 3.8%-4.3%, respectively, were not significantly different between methods. Intraobserver/interobserver agreement for AMS and MS was strong-to-excellent for NFV and RF (ICC ≥0.88). MS demonstrates strongest internal consistency, followed closely by AMS, and AS. Automated segmentation, with or without manual correction, can be considered for 4D flow MRI valvular flow quantification. 3 TECHNICAL EFFICACY: Stage 3.

Left ventricular function assessment including or excluding trabeculations in left ventricular non-compaction, a preliminary case-control cardiac magnetic resonance study.

Functional assessment of compact myocardium and hypertrabeculations in left ventricular non-compaction (LVNC) is underestimated with regards to the morphological spectrum of disease. We aimed to assess whether measuring concurrently left ventricular (LV) volume, mass and ejection fraction (LVEF) with and without trabeculation inclusion on cine magnetic resonance (cineMR) could help diagnose patients with LVNC by comparison to normal individuals with an excess of myocardial trabeculations. This retrospective single center magnetic resonance imaging study (Bichat University Hospital) of 67 consecutive patients with echocardiographic hypertrabeculations seen at echocardiography between March 2011 and October 2018 included 30 patients with known LVNC and 16 control subjects with simple hypertrabeculations (non-compact/compact (NC/C) ratio between 1.8 and 2.2, trabeculations involving 10% to 17% of the left ventricle) using steady-state free precession (SSFP) cine sequences in the standard views. LV volumes, mass and LVEF were measured with and without trabeculation inclusion using CVI42 software. Follow-up was studied in 20 patients and 14 controls. Functional parameters were compared using Student's paired t-test. Pearson product moment correlation coefficients were calculated. Bland-Altman analysis determined the inter- and intra-reader functional data reproducibility. When excluding the trabeculations (i.e. non-compacted myocardium) from measurements, LVEF was within normal ranges both in patients and controls, while it increased by 9.8%±1.6% in LVNC and decreased by 10.9%±1.4% in controls when trabeculae were included in the endocardial contours (P<0.0001). The overall myocardial mass remained stable according to the diastolic or systolic phase in LVNC whereas it significantly decreased in controls. Depending whether trabeculations were included or not, LVEF measurements were significantly different between patients with LVNC and controls. These distinctive measurements might be used as an adjunctive clinical tool to help confirm the diagnosis of LVNC.

Technical note: Minimizing CIED artifacts on a 0.35 T MRI-Linac using deep learning.

Artifacts from implantable cardioverter defibrillators (ICDs) are a challenge to magnetic resonance imaging (MRI)-guided radiotherapy (MRgRT). This study tested an unsupervised generative adversarial network to mitigate ICD artifacts in balanced steady-state free precession (bSSFP) cine MRIs and improve image quality and tracking performance for MRgRT. Fourteen healthy volunteers (Group A) were scanned on a 0.35 T MRI-Linac with and without an MR conditional ICD taped to their left pectoral to simulate an implanted ICD. bSSFP MRI data from 12 of the volunteers were used to train a CycleGAN model to reduce ICD artifacts. The data from the remaining two volunteers were used for testing. In addition, the dataset was reorganized three times using a Leave-One-Out scheme. Tracking metrics [Dice similarity coefficient (DSC), target registration error (TRE), and 95 percentile Hausdorff distance (95% HD)] were evaluated for whole-heart contours. Image quality metrics [normalized root mean square error (nRMSE), peak signal-to-noise ratio (PSNR), and multiscale structural similarity (MS-SSIM) scores] were evaluated. The technique was also tested qualitatively on three additional ICD datasets (Group B) including a patient with an implanted ICD. For the whole-heart contour with CycleGAN reconstruction: 1) Mean DSC rose from 0.910 to 0.935; 2) Mean TRE dropped from 4.488 to 2.877mm; and 3) Mean 95% HD dropped from 10.236 to 7.700mm. For the whole-body slice with CycleGAN reconstruction: 1) Mean nRMSE dropped from 0.644 to 0.420; 2) Mean MS-SSIM rose from 0.779 to 0.819; and 3) Mean PSNR rose from 18.744 to 22.368. The three Group B datasets evaluated qualitatively displayed a reduction in ICD artifacts in the heart. CycleGAN-generated reconstructions significantly improved both tracking and image quality metrics when used to mitigate artifacts from ICDs.

Optimizing Clinical Cardiac MRI Workflow through Single Breath-Hold Compressed Sensing Cine: An Evaluation of Feasibility and Efficiency

Background: Although compressed sensing (CS) accelerated cine holds immense potential to replace conventional cardiovascular magnetic resonance (CMR) cine, how to use CS-based cine appropriately during clinical CMR examinations still needs exploring. Methods: A total of 104 patients (46.5 ± 17.1 years) participated in this prospective study. For each participant, a balanced steady state free precession (bSSFP) cine was acquired as a reference, followed by two CS accelerated cine sequences with identical parameters before and after contrast injection. Lastly, a CS accelerated cine sequence with an increased flip angle was obtained. We subsequently compared scanning time, image quality, and biventricular function parameters between these sequences. Results: All CS cine sequences demonstrated significantly shorter acquisition times compared to bSSFPref cine (p &lt; 0.001). The bSSFPref cine showed higher left ventricular ejection fraction (LVEF) than all CS cine sequences (all p &lt; 0.001), but no significant differences in LVEF were observed among the three CS cine sequences. Additionally, CS cine sequences displayed superior global image quality (p &lt; 0.05) and fewer artifacts than bSSFPref cine (p &lt; 0.005). Unenhanced CS cine and enhanced CS cine with increased flip angle showed higher global image quality than other cine sequences (p &lt; 0.005). Conclusion: Single breath-hold CS cine delivers precise biventricular function parameters and offers a range of benefits including shorter scan time, better global image quality, and diminished motion artifacts. This innovative approach holds great promise in replacing conventional bSSFP cine and optimizing the CMR examination workflow.

Fractal Analysis of Left Ventricular Trabeculae in Patients with End-Stage Renal Disease: A Random Survival Tree Analysis.

The complexity of left ventricular (LV) trabeculae is related to the prognosis of several cardiovascular diseases. To evaluate the prognostic value of LV trabecular complexity in patients with end-stage renal disease (ESRD). Prospective outcome study. 207 participants on maintenance dialysis, divided into development (160 patients from 2 centers) and external validation (47 patients from a third center) cohorts, and 72 healthy controls. 3.0T, steady-state free precession (SSFP) and modified Look-Locker imaging sequences. All participants had their trabecular complexity quantified by fractal analysis using cine SSFP images. Patients were followed up every 2 weeks until April 2023, or endpoint events happened. Random Forest (RF) and Cox regression models including age, diabetes, LV mass index, mean basal fractal dimension (FD), and left atrial volume index, were developed to predict major adverse cardiac events (MACE). Patients were divided into low- and high-risk groups based on scores derived from the RF model and survival compared. Receiver operating characteristic curve analysis; Kaplan-Meier survival analysis with log rank tests; Harrel's C-index to assess model performance. A P value <0.05 was considered statistically significant. Fifty-five patients (26.57%) experienced MACE during a median follow-up time of 21.83 months. An increased mean basal FD (≥1.324) was associated with a significantly higher risk of MACE. The RF model (C-index: 0.81) had significantly better discrimination than the Cox regression model (C-index: 0.74). Participants of the external validation dataset classified into the high-risk group had a hazard of experiencing MACE increased by 12.29 times compared to those in the low-risk group. LV basal FD was an independent predictor for MACE in patients with ESRD. Reliable risk stratification models could be generated based on LV basal FD and other MRI variables using RF analysis. 2 TECHNICAL EFFICACY: Stage 2.

Utilizing Artificial Intelligence-Based Deformable Registration for Global and Layer-Specific Cardiac MRI Strain Analysis in Healthy Children and Young Adults

The absence of published reference values for multilayer-specific strain measurement using cardiac magnetic resonance (CMR) in young healthy individuals limits its use. This study aimed to establish normal global and layer-specific strain values in healthy children and young adults using a deformable registration algorithm (DRA). A retrospective study included 131 healthy children and young adults (62 males and 69 females) with a mean age of 16.6±3.9 years. CMR examinations were conducted using 1.5T scanners, and strain analysis was performed using TrufiStrain research prototype software (Siemens Healthineers, Erlangen, Germany). Global and layer-specific strain parameters were extracted from balanced Steady-state free precession cine images. Statistical analyses were conducted to evaluate the impact of demographic variables on strain measurements. The peak global longitudinal strain (LS) was -16.0±3.0%, peak global radial strain (RS) was 29.9±6.3%, and peak global circumferential strain (CS) was -17.0±1.8%. Global LS differed significantly between males and females. Transmural strain analysis showed a consistent pattern of decreasing LS and CS from endocardium to epicardium, while radial strain increased. Basal-to-apical strain distribution exhibited decreasing LS and increasing CS in both global and layer-specific analysis. This study uses DRA to provide reference values for global and layer-specific strain in healthy children and young adults. The study highlights the impact of sex and age on LS and body mass index on RS. These insights are vital for future cardiac assessments in children, particularly for early detection of heart diseases.

Prognostic Significance of Biventricular and Biatrial Strain in Dilated Cardiomyopathy: Strain Analysis Derived from Cardiovascular Magnetic Resonance.

Dilated cardiomyopathy (DCM) has a poor prognosis and high mortality. The relationship between the deformation capacity of the biatrial and biventricular regions in patients with DCM remains unclear. This retrospective study used cardiovascular magnetic resonance (CMR) to assess patient enrollment between September 2020 to May 2022. Feature tracking (FT) was used to evaluate biventricular global radial strain (GRS), global circumferential strain (GCS) and global longitudinal strain (GLS). Fast long-axis method was used to evaluate biatrial GLS by analyzing balanced steady-state free precession cine images. The median follow-up period was 362 days (interquartile range: 234 to 500 days). DCM patients were divided into two groups based on the occurrence or non-occurrence of major adverse cardiac event (MACE). The primary endpoint was defined as all-cause death, heart transplantation, and adverse ventricular arrhythmia. The secondary end point included hospitalizations due to heart failure. Cox regression analysis was utilized for variables and Kaplan-Meier survival was utilized for clinical outcomes. There were 124 DCM patients (52.82 12.59 years, 67.74% male) and 53 healthy volunteers (53.17 14.67 years, 52.83% male) recruited in this study. Biventricular GRS, GCS, GLS, and biatrial GLS were significantly impaired in the DCM group compared with the healthy group. In receiver-operating characteristic curve, biatrial GLS and biventricular GRS, GCS, and GLS showed significant prognostic value in predicting MACEs (all p 0.05). In multivariate Cox regression analysis, left ventricular (LV) GLS offered a significant and independent prognostic value surpassing other CMR parameters in predicting MACE. In Kaplan-Meier analysis, patients with a LV GLS -4.81% had a significantly higher rate of MACE (Log-rank p 0.001). LV GLS was independently associated with MACEs in DCM patients by using FT and fast long-axis method derived from CMR. Comprehensive CMR examination including biatrial and biventricular functions should be systematically performed, to understand disease characteristics, as well as improve the risk stratification and therapeutic management for patients with DCM.

Aligning Multi-Sequence CMR Towards Fully Automated Myocardial Pathology Segmentation.

Myocardial pathology segmentation (MyoPS) is critical for the risk stratification and treatment planning of myocardial infarction (MI). Multi-sequence cardiac magnetic resonance (MS-CMR) images can provide valuable information. For instance, balanced steady-state free precession cine sequences present clear anatomical boundaries, while late gadolinium enhancement and T2-weighted CMR sequences visualize myocardial scar and edema of MI, respectively. Existing methods usually fuse anatomical and pathological information from different CMR sequences for MyoPS, but assume that these images have been spatially aligned. However, MS-CMR images are usually unaligned due to the respiratory motions in clinical practices, which poses additional challenges for MyoPS. This work presents an automatic MyoPS framework for unaligned MS-CMR images. Specifically, we design a combined computing model for simultaneous image registration and information fusion, which aggregates multi-sequence features into a common space to extract anatomical structures (i.e., myocardium). Consequently, we can highlight the informative regions in the common space via the extracted myocardium to improve MyoPS performance, considering the spatial relationship between myocardial pathologies and myocardium. Experiments on a private MS-CMR dataset and a public dataset from the MYOPS2020 challenge show that our framework could achieve promising performance for fully automatic MyoPS.

MRI Quantification of Left Atrial Circumferential Strain in Mitral Regurgitation: A Feasibility and Reproducibility Study.

MRI-derived left atrial (LA) longitudinal strain has been shown to be a marker for mitral regurgitation, but the utility of LA circumferential strain remains unclear. To assess feasibility and reproducibility of LA circumferential strain, identify changes in mitral regurgitation patients compared to healthy volunteers, and determine strain's association with mitral regurgitation severity and cardiac function. Retrospective. 52 mitral regurgitation patients, 12 healthy volunteers. Steady-state free precession cine and 2D phase contrast sequences at 1.5T. Peak LA circumferential strain was computed in each of three short-axis slices (superior, mid, and inferior) and longitudinal strain computed from long-axis slices using MRI feature-tracking software. Strain test-retest reproducibility was determined from two repeat studies in healthy volunteers. LA circumferential strain test-retest reproducibility was assessed by intra-class correlation coefficient (ICC). Strain was compared between cohorts using Student's t-test or Mann-Whitney U. Mitral regurgitation severity association with strain and LV function was assessed using Spearman correlation and multivariable regression. Significance was defined as P < 0.05. LA circumferential strain assessment was feasible in all subjects with moderate reproducibility in the superior (ICC = 0.74), mid LA (ICC = 0.71), and inferior LA (ICC = 0.63). In mitral regurgitation patients, LA circumferential strain was significantly lower in the superior (11.86% [6.5%,19.2%] vs. 18.73% ± 6.7%) and mid LA slices (18.41% ± 9.5% vs. 28.7% ± 10.4%) compared to healthy volunteers. Mitral regurgitation severity significantly associated with mid LA circumferential strain (β = -0.03) and LAV significantly associated with superior LA circumferential strain (β = -2.09), both independent of LA longitudinal strain and CO. LA circumferential strain assessment is feasible with moderate reproducibility. Compared to healthy volunteers, patients had significantly lower LA circumferential strain. Mitral regurgitation severity and LAV were significantly associated with LA circumferential strain independent of LA longitudinal strain. 3 TECHNICAL EFFICACY: Stage 2.

Epicardial fat is associated with adverse ventricular remodeling and diastolic dysfunction in women with ischemia but no obstructive coronary artery disease

Abstract Background Epicardial fat is a metabolically active adipose tissue depot located between the myocardium and visceral pericardium. Increased epicardial fat is associated with the development of atherosclerosis and obstructive coronary artery disease (CAD). However, information regarding relations between epicardial fat—which covers 80% of the heart surface—and cardiac function, is limited. Purpose To evaluate relations between epicardial fat and cardiac morphology and function in all four chambers of the heart. Methods To accomplish our objective, we studied 113 participants from the Women’s Ischemia Syndrome Evaluation cohort who underwent invasive measurement of left ventricular end-diastolic pressure and coronary reactivity, and comprehensive rest-stress cardiac magnetic resonance imaging (cMRI) per standardized protocols. Epicardial fat area was measured from a single high resolution steady-state free precession cine cMRI in the horizontal long axis plane (Figure 1A). Left ventricular (LV) mass, bi-ventricular volumes, and ejection fractions, were assessed using a stack of short axis cine images covering both ventricles (Figure 1B). LV concentricity was defined as the mass-to-end-diastolic volume ratio. Strain and strain rate were assessed in all four chambers by feature tracking (Figure 1C). Left atrial (LA) volume was assessed using three orthogonal long-axis cines, while right atrial (RA) area was assessed using a single horizontal long axis cine (Figure 1D). Participants were stratified by the median epicardial fat area, defined as "Low" epicardial fat (&amp;lt;13 cm2, n=48) and "High" epicardial fat (≥13 cm2, n=65). Results Participant characteristics are reported in Table 1. Notably, the majority (58%) had "High" epicardial fat and body mass index (BMI) was greater in the "High" epicardial fat group. After adjusting for BMI, however, bi-ventricular stroke index remained lower in the "High" epicardial fat group, as did bi-atrial ejection fraction (fractional area change, RA). Moreover, compared to the "Low" epicardial fat group, LV concentricity and LA booster strain were higher in the "High" epicardial fat group (Figure E-F), while LV longitudinal strain and LV early circumferential diastolic strain rate were lower (Figure G-H). No other major differences were observed between groups, including coronary reactivity or myocardial perfusion reserve. Conclusions Together, these data suggest that higher BMI is associated with increased epicardial fat that is related to adverse cardiac remodeling, along with systolic and diastolic dysfunction, across multiple cardiac chambers. Further studies would be useful to define specific mechanism contributing to these observations.Figure 1Table 1